Hyperspectral image classification using support vector machines (SVM) after wavelet domain denoising is proposed in this paper. In the proposed approach, hyperspectral images are classified using SVM after noise reduction is carried out in each band independent of other bands using spatially adaptive Bayesian shrinkage. It is shown that support vector machine classification of denoised hyperspectral images gives significantly better classification accuracy and furthermore improves sparsity. Therefore this approach has faster testing time, compared with direct SVM based classification. This feature makes the denoised SVM based hyperspectral classification approach more suitable for applications that require low-complexity, and possibly real-time classification.